I think it's a little unfair to pretend the quadratic range term dominates the "linear" density term. If you think about it for a moment, the density describes

I think it's a little unfair to pretend the quadratic range term dominates the "linear" density term. If you think about it for a moment, the density describes area, so already has a ^-2 (to the -2 power) incorporated into it. If we lived in a 3D world, there would be a negative third power incorporated into it (although you'd also have a range^3 term) In reality both terms contribute equally. Considering the range of densities and ranges available in the USA, density currently makes more of an impact. Speeds range from 10mph at the minimum (a typical net average speed if you're biking or taking subways and walking connections through Manhattan) to 60mph at the most (a typical speed limit on a rural highway) That's about one decimal order of magnitude. Density on the other hand varies much, much more. There are rural areas with a handful of people in a square mile, and places with 50,000 or more people in a square mile. That's more like 4 orders of magnitude. This matches up with the anecdotal observation that in terms of "How many employers/restaurants/people/etc can I get to within a 15 minute trip" is far larger in denser bigger cities, despite travel times being slower. Faster travel times just cannot keep up with the low density needed to allow cars to reach those speeds.